Skip to contentSkip to conversation

netfilter: replace skb_try_make_writable() by skb_ensure_writable()

PATCH Superseded Pablo Neira Ayuso 1 participant 3 messages Apr 27

Subsystems (2)

v5 of 7 Superseded Latest is v7 →

lore
Share & tools
lore

Pablo Neira Ayuso

author maintainer
#0 lore raw

PATCH v5 1/3 netfilter: replace skb_try_make_writable() by skb_ensure_writable()

skb_try_make_writable() only works on clones and uncloned packets might have their network header in paged fragments.

nft_fwd needs to work for the ingress and egress hooks, but the egress hook where skb->data points to the mac header, use skb_network_offset() to include the mac header. The flowtable is fine since it already uses the transport offset.

Fixes d32de98ea70f ("netfilter: nft_fwd_netdev: allow to forward packets via neighbour layer")
Fixes 7d2086871762 ("netfilter: nf_flow_table: move ipv4 offload hook code to nf_flow_table")

---
v5: no changes

 net/netfilter/nf_flow_table_ip.c | 4 ++--
 net/netfilter/nft_fwd_netdev.c   | 5 +++--
 2 files changed, 5 insertions(+), 4 deletions(-)
net/netfilter/nf_flow_table_ip.c +2 -2
index fd56d663cb5b..dbd7644fdbeb 100644
@@ -524,7 +524,7 @@ static int nf_flow_offload_forward(struct nf_flowtable_ctx *ctx,
 		return 0;
 	}
 
-	if (skb_try_make_writable(skb, thoff + ctx->hdrsize))
+	if (skb_ensure_writable(skb, thoff + ctx->hdrsize))
 		return -1;
 
 	flow_offload_refresh(flow_table, flow, false);
@@ -1037,7 +1037,7 @@ static int nf_flow_offload_ipv6_forward(struct nf_flowtable_ctx *ctx,
 		return 0;
 	}
 
-	if (skb_try_make_writable(skb, thoff + ctx->hdrsize))
+	if (skb_ensure_writable(skb, thoff + ctx->hdrsize))
 		return -1;
 
 	flow_offload_refresh(flow_table, flow, false);
net/netfilter/nft_fwd_netdev.c +3 -2
index 4bce36c3a6a0..2cc809303ce8 100644
@@ -100,6 +100,7 @@ static void nft_fwd_neigh_eval(const struct nft_expr *expr,
 	int oif = regs->data[priv->sreg_dev];
 	unsigned int verdict = NF_STOLEN;
 	struct sk_buff *skb = pkt->skb;
+	int nhoff = skb_network_offset(skb);
 	struct net_device *dev;
 	int neigh_table;
 
@@ -111,7 +112,7 @@ static void nft_fwd_neigh_eval(const struct nft_expr *expr,
 			verdict = NFT_BREAK;
 			goto out;
 		}
-		if (skb_try_make_writable(skb, sizeof(*iph))) {
+		if (skb_ensure_writable(skb, nhoff + sizeof(*iph))) {
 			verdict = NF_DROP;
 			goto out;
 		}
@@ -132,7 +133,7 @@ static void nft_fwd_neigh_eval(const struct nft_expr *expr,
 			verdict = NFT_BREAK;
 			goto out;
 		}
-		if (skb_try_make_writable(skb, sizeof(*ip6h))) {
+		if (skb_ensure_writable(skb, nhoff + sizeof(*ip6h))) {
 			verdict = NF_DROP;
 			goto out;
 		}
Signature · 1 line
2.47.3
author maintainer
#1 lore raw

PATCH v5 2/3 netfilter: nft_fwd_netdev: add device and headroom validate with neigh forwarding

The ttl field has been decremented already and evaluation of this rule would proceed, just drop this packet instead if there is no destination device to forwards this packet. This is exactly what nf_dup already does in this case.

Moreover, use skb_cow_head() to ensure there is sufficient headroom when forwarding this via neigh_xmit().

Fixes d32de98ea70f ("netfilter: nft_fwd_netdev: allow to forward packets via neighbour layer")

---
v5: rewrite patch title and add skb_cow_head() call

 net/netfilter/nft_fwd_netdev.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
net/netfilter/nft_fwd_netdev.c +9 -2
index 2cc809303ce8..94d6397f10a8 100644
@@ -153,8 +153,15 @@ static void nft_fwd_neigh_eval(const struct nft_expr *expr,
 	}
 
 	dev = dev_get_by_index_rcu(nft_net(pkt), oif);
-	if (dev == NULL)
-		return;
+	if (dev == NULL) {
+		verdict = NF_DROP;
+		goto out;
+	}
+
+	if (skb_cow_head(skb, LL_RESERVED_SPACE(dev))) {
+		verdict = NF_DROP;
+		goto out;
+	}
 
 	skb->dev = dev;
 	skb_clear_tstamp(skb);
Signature · 1 line
2.47.3
author maintainer
#2 lore raw

PATCH v5 3/3 netfilter: nft_fwd_netdev: use recursion counter in neigh egress path

In reply to Pablo Neira Ayuso

From: Weiming Shi <[email protected]>

nft_fwd_neigh can be used in egress chains (NF_NETDEV_EGRESS). When the forwarding rule targets the same device or two devices forward to each other, neigh_xmit() triggers dev_queue_xmit() which re-enters nf_hook_egress(), causing infinite recursion and stack overflow.

Move the nf_get_nf_dup_skb_recursion() accessor and NF_RECURSION_LIMIT to the shared header nf_dup_netdev.h as a static inline, so that nft_fwd_netdev can use the recursion counter directly without exported function call overhead. Guard neigh_xmit() with the same recursion limit already used in nf_do_netdev_egress().

[ Updated to cache the nf_get_nf_dup_skb_recursion pointer. --pablo ]

Fixes f87b9464d152 ("netfilter: nft_fwd_netdev: Support egress hook")

---
v5: no changes

 include/net/netfilter/nf_dup_netdev.h | 13 +++++++++++++
 net/netfilter/nf_dup_netdev.c         | 16 ----------------
 net/netfilter/nft_fwd_netdev.c        |  8 ++++++++
 3 files changed, 21 insertions(+), 16 deletions(-)
include/net/netfilter/nf_dup_netdev.h +13 -0
index b175d271aec9..609bcf422a9b 100644
@@ -3,10 +3,23 @@
 #define _NF_DUP_NETDEV_H_
 
 #include <net/netfilter/nf_tables.h>
+#include <linux/netdevice.h>
+#include <linux/sched.h>
 
 void nf_dup_netdev_egress(const struct nft_pktinfo *pkt, int oif);
 void nf_fwd_netdev_egress(const struct nft_pktinfo *pkt, int oif);
 
+#define NF_RECURSION_LIMIT	2
+
+static inline u8 *nf_get_nf_dup_skb_recursion(void)
+{
+#ifndef CONFIG_PREEMPT_RT
+	return this_cpu_ptr(&softnet_data.xmit.nf_dup_skb_recursion);
+#else
+	return &current->net_xmit.nf_dup_skb_recursion;
+#endif
+}
+
 struct nft_offload_ctx;
 struct nft_flow_rule;
 
net/netfilter/nf_dup_netdev.c +0 -16
index e348fb90b8dc..3b0a70e154cd 100644
@@ -13,22 +13,6 @@
 #include <net/netfilter/nf_tables_offload.h>
 #include <net/netfilter/nf_dup_netdev.h>
 
-#define NF_RECURSION_LIMIT	2
-
-#ifndef CONFIG_PREEMPT_RT
-static u8 *nf_get_nf_dup_skb_recursion(void)
-{
-	return this_cpu_ptr(&softnet_data.xmit.nf_dup_skb_recursion);
-}
-#else
-
-static u8 *nf_get_nf_dup_skb_recursion(void)
-{
-	return &current->net_xmit.nf_dup_skb_recursion;
-}
-
-#endif
-
 static void nf_do_netdev_egress(struct sk_buff *skb, struct net_device *dev,
 				enum nf_dev_hooks hook)
 {
net/netfilter/nft_fwd_netdev.c +8 -0
index 94d6397f10a8..2c4abf24f26a 100644
@@ -95,6 +95,7 @@ static void nft_fwd_neigh_eval(const struct nft_expr *expr,
 			      struct nft_regs *regs,
 			      const struct nft_pktinfo *pkt)
 {
+	u8 *nf_dup_skb_recursion = nf_get_nf_dup_skb_recursion();
 	struct nft_fwd_neigh *priv = nft_expr_priv(expr);
 	void *addr = &regs->data[priv->sreg_addr];
 	int oif = regs->data[priv->sreg_dev];
@@ -152,6 +153,11 @@ static void nft_fwd_neigh_eval(const struct nft_expr *expr,
 		goto out;
 	}
 
+	if (*nf_dup_skb_recursion > NF_RECURSION_LIMIT) {
+		verdict = NF_DROP;
+		goto out;
+	}
+
 	dev = dev_get_by_index_rcu(nft_net(pkt), oif);
 	if (dev == NULL) {
 		verdict = NF_DROP;
@@ -165,7 +171,9 @@ static void nft_fwd_neigh_eval(const struct nft_expr *expr,
 
 	skb->dev = dev;
 	skb_clear_tstamp(skb);
+	(*nf_dup_skb_recursion)++;
 	neigh_xmit(neigh_table, dev, addr, skb);
+	(*nf_dup_skb_recursion)--;
 out:
 	regs->verdict.code = verdict;
 }
Signature · 1 line
2.47.3